Countering the Next Generation of Terrorists: Rule of Law and Long-Term Counterterrorism Strategy

COUNTERING THE NEXT GENERATION OF TERRORISTS: RULE OF LAW AND LONG-TERM COUNTERTERRORISM STRATEGY

A Thesis submitted to the Faculty of the Graduate School of Arts and Sciences of Georgetown University in partial fulfillment of the requirements for the degree of Master of Public Policy in Public Policy

Ryan B. Greer, B.A.

Washington, DC April 15, 2011

COUNTERING THE NEXT GENERATION OF TERRORISTS: RULE OF LAW AND LONG-TERM COUNTERTERRORISM STRATEGY

Ryan B. Greer, B.A.

Thesis Advisor: Matthew H. Fleming, PhD

ABSTRACT

Counterterrorism efforts in the short term are comprised of military and law enforcement

activity (not least those focused on capturing or killing malicious actors). However, in the long

term, such activity may not be sufficient to reduce the creation of the next generation of terrorists, and efforts must be made to prevent individuals from turning to terrorism in the first place. Previous analyses and discussions of this subject have suggested that individuals turn to terrorism because of factors such as impediments to democratic institutions or poor economic conditions. This paper analyzes the country of origin of terrorists to test the hypothesis that rule of law contributes to the development of terrorist proclivities and finds that weak rule of law is a contributing factor in generating terrorists. This paper also discusses the implications for resource allocation in a post-conflict reconstruction environment, where rule of law conditions are more malleable and counterterrorism represents a salient interest.

iii

Acknowledgements

I would like to thank my thesis advisor, Dr. Matthew Fleming for his invaluable assistance throughout this process. I also greatly appreciate the input of Dr. Jean Mitchell,

Lieutenant Colonel John Gossart, Michael Barker and Christopher Melenovsky. This paper represents only the views of the author, but it would not be what it is without their sharp minds and selfless assistance.

Table of Contents

Abstract iii

Acknowledgments iv

Introduction 1

Chapter 1. Background 3 Rule of Law and Social Conditions Prone to Terrorism 3 This Study: Rule of Law for Long-Term Counterterrorism 4 Assessing Terrorism-Prone States and Possible Motivations For Terrorism 4 Good Governance and Rule of Law 7 Review of Related Literature 8

Chapter 2. Analysis 17 Conceptual Framework 17 Data Sources 20 Analysis Methodology 29 Findings 31 Interpretation of Results 34

Chapter 3. Policy Implications and Recommendations 36 Assessing At-Risk Countries 36 Changing Priorities and Shifting Resources 37 Case Study: Post-Conflict Reconstruction in Afghanistan 38 Strategy for Rebuilding Afghanistan 38 Safeguarding Individual Rights: Afghanistan’s Women 39 Enforcing the Law: Money Laundering in Afghanistan 41 Policy Lessons: Linking Short-and Long-Term Counterterrorism Objectives 42 Conclusions 43

Appendix: Regression Models 45

References 47

Table of figures

Figures 17 Basic Counterterrorism Strategies 17 Conceptual Framework for Terrorist Motivations 18

Descriptive Statistics 23 Dependent Variable: Terrorist Perpetrators by Nationality 23 Explanatory Variable: Individual Rights 24 Explanatory Variable: Commercial Law 25 Explanatory Variable: Law Enforcement 27 Explanatory Variable: Similarity to Democracy 28 Summary 29

Findings 32 Collinearity Matrix 33 Findings: Details 45 Robustness Check: Tobit 45 Outlier Correction 46

Introduction

Individuals are not born terrorists; rather they are radicalized in some way. Given that some countries are more likely to be the country of origin for terrorists, it is reasonable to question whether there are certain structural conditions that contribute to the likelihood that a population will be radicalized. In common public discourse, there are claims that many different motivations drive the radicalization of individuals who ultimately turn to terrorism. The most commonly mentioned motivations include an anarchic international community, United States

(U.S.) hegemony, lack of domestic democracy, manipulation of religion, and/or economic factors

(Thayer and Hudson, 2010). Some of these motivations can be addressed by policy shifts and efforts by the U.S. and others to mitigate factors that may influence the decisions of individuals to turn to terrorism; but many motivations, such as religion, have no such policy lever for foreign-policymakers to pull when attempting to reduce the number of people who turn to terrorism.

This paper examines structural conditions that can increase the likelihood that individuals will turn to terrorism and upon which the U.S. may have an impact through development or post- conflict reconstruction. This paper gives particular attention to rule of law factors as they compare to economic and political factors and seeks to answer the following question: Are state- level rule of law conditions significant predictors of terrorist origin? Then, upon finding that rule of law does impact whether terrorists originate a country, this paper offers some policy adjustments that may strengthen rule of law in countries whose populations may be at-risk for terrorist radicalization.

The next section of this paper (Chapter 1) discusses the background of the foreign policy scope of international terrorism with respect to the conditions associated with countries prone to

terrorism. Chapter 1 also discusses why policy questions such as these are important and what

types of information are typically analyzed and why, reviewing related literature and elucidating

the motivations for approaching the topic in this way. The subsequent section (Chapter 2)

explains the quantitative analysis, including the limitations to the data used and interpretations of

results. The final section (Chapter 3) explains some of the possible policy responses given the

results of the analysis. Due to the breadth of possible policy approaches and greater relevance

gleaned from concrete examples, the section on policy implications is focused on post-conflict reconstruction in Afghanistan. In this way, this paper seeks to analyze the countries whose populations are vulnerable to radicalizing as terrorists and seeks to explain the policy responses that may address the relevant problems with respect to a specific example to a particularly at-risk

country.

Chapter 1. Background

1.1 Rule of Law and Social Conditions Prone to Terrorism

As noted, this study focuses on rule of law as it compares to other factors. Rule of law itself has various definitions across related literature. Susan Rose-Ackerman defines “rule of law” as “legal limits, both civil and criminal, on private interactions” and “limits on the political regime,” suggesting that rule of law encompasses law enforcement, dispute resolution, and protection of individual liberties (Rose-Ackerman, 2004). Countries that provide the means and incentives to enforce laws and provide the foundation for the legal protection of individual rights not only provide outlets for dispute resolution, but engender social norms conducive to respect for order.

In relation to social norms, rule of law is undermined in societies in which informal trials that lack an adequate evidentiary standard and the authority of a legitimate government dictate justice. Lack of adherence to standardized laws and procedures not only breeds oppression but engenders social norms accepting of illegitimate forms of justice (e.g., physical retribution for perceived injustices, rather than standardized trials).

In addition, lack of a formal dispute resolution mechanism or strong property rights statutes can have similar impacts on social behavior. If disputes cannot be resolved in an impartial judicial system, people may begin to look for retribution outside of that system.

Moreover, if property is centralized in the hands of the privileged through a ruling regime, and there is no means for the less well-off to improve their socioeconomic status, resentment can ensue. Therefore, the capacity to lend or declare bankruptcy could encourage social mobility and reduce demand for extra-judicial retribution due to commercial disputes.

Rule of law also has significant implications for criminal law and corruption. Law

enforcement entities play a large role in preventing the organizational capacity of radicalized

groups, and can reduce corruption that may engender social norms distrustful of government or

commerce. Money laundering statutes and the means to enforce them can mitigate and deter the primary means by which organized criminals build organizational capacity and by which illicit groups obfuscate their functions.

1.2 This Study: Rule of Law for Long-Term Counterterrorism

These examples illustrate the capacity of law enforcement, protection of individual rights,

and formalized dispute resolution to reduce social norms and legal vulnerabilities that may foster

societies conducive to, or enabling of, terrorism. This paper seeks to test whether societies that embrace law enforcement, individual rights, and dispute resolution are less likely to have terrorist perpetrators originate in them. This paper could be considered an evaluation of priorities for international diplomacy and development strategies in their roles in counterterrorism by assessing conditions most likely to lead to terrorist radicalization. By comparing economic

indicators to conditions such as the protection of individual rights, conclusions can be drawn for

prioritization in resource allocation. As such, post-conflict reconstruction in Afghanistan presents

itself as an appropriate case study, as it has historically been a terrorism-enabling country, and

the U.S. resources provided for rebuilding it could shape the legal, economic, and governance

mechanisms and successes.

1.3 Assessing Terrorism-Prone States and Possible Motivations for Terrorism

In order to examine whether policy can reduce the risk of terrorist development, it is necessary to determine the common conditions in states producing terrorists. A macro-level

4 structural assessment can help policymakers frame issues, assess priorities, and clarify broad circumstances to reform counterterrorism policy.

Varying structural conditions across different countries can create environments more or less conducive to the development of terrorism. These structural factors are commonly explained with respect to the motivations they create for individuals to turn to terrorism. First, an anarchic international system may lead to transnational angst when groups feel they are under-recognized by the international community, particularly with the lack of a police force or governance mechanism at the international level. International bodies such as the United Nations (UN), while providing a means for international states to organize, do not govern; instead, most dispute resolution negotiations on the international stage involve a more powerful stakeholder (e.g., the

U.S.) having excess bargaining power. The perception of this excess bargaining power could lead groups to embrace asymmetric warfare in order to level the playing field. One method to mitigate this factor could include international recognition of legitimacy for groups that agree to renounce violence. Recognition of legitimacy could create from the terrorist group a new stakeholder in the international community, which provides an incentive to renounce terrorism; however, this would require affirming the actions of terrorists, which, aside from being imprudent itself, could cause others to mimic their tactics in the hopes of a payout.

Second, the global hegemony of the U.S. makes it a target for groups hoping to create a power vacuum that they may fill, and creates a perception of colonization-like dominance over developing countries dependent on ties to the U.S. Solutions for this as a cause of terrorist activity may include “strategic communication” to improve the perception of the U.S. (such as including an American flag symbol on crates of food feeding the poor in rural Afghanistan).

However, policies to mitigate the frustration of U.S. hegemony are limited. The U.S. could not

be expected to limit its global influence, nor is it practical to assume the U.S. can communicate its benevolence adequately enough to quell perception of its undue dominance, and with it, the threat of terrorist tactics in response.

Third, manipulation of religion (Islam is typically discussed in this context) could be used

– regardless of the motivation for its use – to convince potential terrorist recruits that it is an

imperative of their faith to engage in terrorism. While religion may certainly be a prevalent tool

for recruiters, there is little that can be done to stop religious speech, outside of outreach and

communication of countervailing views, which are difficult and unreliable. Moreover, while

religion may be a motivation to attack a potential target, such as the U.S., it cannot explain the

tactics. That is, religion may call for a desired result, but the use of terrorist tactics results from

the individual’s choice to engage in them.

Fourth, economic stagnation may both foster the angst associated with radicalized

violence and also allow for financial incentives to have influence over potential recruits. Poverty

leads to social angst and the U.S. can be perceived as a target of poverty-related angst because

the U.S. thrives in an international system that is perceived as failing to provide for the

impoverished. Moreover, those who are unemployed are more likely to have the time to plan

terrorist activities, less to lose from engaging in them, and more to gain from an illicit

environment generally (and many terrorist groups offer financial incentives, such as funding for

families of suicide bombers). Strategies to mitigate this problem include economic development

and financial ties between at-risk states (those ranking high on a terrorism risk assessment or

showing signs of state failure) and politically-targeted states (such as the U.S.).

Lastly, lack of democracy may drive those without a voice in the political process to

attempt to gain that voice through violence. However, democratic elections can elect oppressive

regimes or even terrorists. Moreover, it is unlikely to be possible to engineer the transition to a

democratic regime from an autocratic one. In fact, regime change may never be possible, as

evident by the questionable successes in Iraq and Afghanistan, particularly in light of the

resource drain associated with coercive regime change. Instead, the conclusion we may draw

from the assumption of this causal link between terrorism and a lack of democracy is that the

principles of good governance typically present in a democratic country may be more relevant

than democracy as a method of legitimizing governance.

1.4 Good Governance and Rule of Law

For a country to achieve stability and order through good governance, a strong legal

system and adherence to the law is paramount. Recognizing this, international development and

post-conflict reconstruction strategies frequently prioritize rule of law to encourage sustainable

stability. According to Rosa Ehrenreich Brooks, Deputy Assistant Secretary of Defense for Rule

of Law and Humanitarian Policy, “The war on terrorism has given further impetus to U.S. and

international enthusiasm for rule-of-law promotion; most recently, millions have been pledged

for rule-of-law programs in post-Taliban Afghanistan and post-war Iraq,” but “the formal

dimensions of the rule of law” do not always “lead reliably and predictably to the emergence of a

robust societal commitment to the more substantive aspects of the rule of law” (Brooks, 2003).

The logical next question would then be: what conditions must follow the creation of rule of law institutions that foster social norms that reduce terrorism? It is possible that creating stakeholders in rule of law through commercial incentives and redress of violations of individual rights will assist in shifting social norms, and that creating legal and law enforcement institutions for competing stakeholders to enforce checks and balances will broaden the appeal of stable governance. This implies that an accurate assessment of rule of law focuses on outcomes and

policy responses on stakeholder creation, rather than on the presence or absence of legal

institutions.

The objective of this paper is to determine the societal conditions that have the greatest impact on creating environments conducive to terrorism, with particular respect given to the strength of rule of law. While it is impossible to force a state to change the religion of its people, and highly implausible to impose democracy, it is possible to assist in supporting strong rule of law in a manner that will both take the society’s norms into account and also encourage sustainable stability.

1.5 Review of Related Literature

As the salient conflicts in which the world’s only superpower is engaged have shifted from those of state-to-state combat to those against transnational terrorist groups, much of counterterrorism study has been motivated by prior research of state-based problems and conflicts. The lessons learned from previous iterations of conflict analysis can help frame the issue of counterterrorism strategy.

One explanation for violent conflict, including terrorism, is state failure; as such, state failure assessments can provide meaningful insight into conditions that foster terrorism. Risk assessments for state failure “focus monitoring and analytical attention to high-risk situations before they are fully developed and provide a framework for interpreting the results of real-time monitoring” (Carment, 2004). As David Carment writes in his piece “Preventing State Failure,” risk assessment models use either configurational variables (“interaction among units of analysis,” e.g., warring factions) or composite variables (“the attributes and performance indicators of the state over time”) in order to provide “early warning about the risk of state failure” (Carment, 2004). As such, risk assessments use “country performance measures for the

factors that directly and indirectly influence a country’s security and level of stability” (Carment,

2004). Carment criticizes the dominant literature in this area for failing to provide “policy-

relevant diagnosis” because, while studying conflict for educational purposes can be helpful, the

“ultimate goal is to provide risk assessments in a way that is actually useful to decision makers

so they can better respond to potential, emerging or existing risks” (Carment, 2004). Carment

identifies studies in this area that provide (i) macro-level evaluation of structural indicators

(“econometrically or through pattern recognition techniques”); (ii) macro-level time series of

leading indicators; (iii) intermediate-level conjunctural models (that “track changes in pre-

specified events and interactions among groups);” (iv) intermediate-level structured and

subjective models where experts identify key actors and “estimate their future position…with

regard to their power to influence the outcome” of a conflict; (v) micro-level sequential models that track specific behaviors; (vi) micro-level response models that evaluate outside response to conflict “and develop feasibility assessments based therein;” and (vii) micro-level field reporting

(Carment, 2004). Micro-level assessments tend to be limited in the quality and quantity of data, particularly in terms of external validity for international-level studies. Additionally, allocating resources to field reporting to create micro-level analyses must be based first on an indicator of risk from a macro-level assessment. Therefore, a structural analysis must precede field-based analysis and be geared toward informing policy.

This paper analyzes terrorism as a macro-level time series evaluation of structural indicators using composite variables, making it a precursor to micro-level studies or as a motivator for broad policy shift. As Carment notes, “structural approaches emphasize capacity building to provide conflict-prone societies with a means to address root causes of conflict,” and as such, this paper focuses on addressing the root social causes of terrorism (Carment, 2004).

In order to determine the conditions that may contribute to terrorism, it is necessary to

understand some of the structural factors that contribute to the logic of becoming a terrorist. In

“The Strategic Logic of Suicide Terrorism,” Robert Pape discusses the idea of suicide terrorism

falling into “rational irrationality” (Pape, 2004). He lays out many of the coercive and targeting

strategies that terrorist recruiters utilize to accomplish the goal of incentivizing and channeling

those who may be willing to commit a suicide attack. Pape points out that, while previous

psychological profiles of terrorists focused on “uneducated, unemployed, socially isolated, single

men in their late teens and early 20s,” contemporary terrorists come from increasingly different

backgrounds representing disparate upbringings, age ranges, ethnicities, and genders (Pape,

2004). Pape’s analysis is aimed toward the cyclical nature of suicide terrorism as a tactic, noting

cases that may exhibit trends, rather than comparing structural factors in societies. That is,

Pape’s focus is on the means and opportunity of committing an illicit act, rather than the structural trends associated with the intent of committing it. Therefore, his study’s value begins

after the objectives of the perpetrator have been conceived.

Similarly, in “Terrorist Factions,” Ethan Bueno de Mesquita looks at the conditions that

impact the likelihood of a splinter faction forming once a terrorist faction has been terminated.

This may be useful to motivate the metrics by which we can assess the likelihood of defeating a

terrorist organization, as opposed to, say, killing certain members only to give rise to an equally

or more dangerous faction. While Bueno de Mesquita’s analysis of outlets for legal grievance

and their impact on the motivation to form an additional group may contribute to the conceptual

basis for rule of law as a counterterrorism strategy, the analysis is micro-level and is premised on

the assumption that there is already a faction formed.

In “Economic Conditions and the Quality of Suicide Terrorism,” Benmelech et al. (2010) lay out the economic theory of civilian casualties in suicide attacks. This study compares economic variables with demographic and security variables in assessing their impact on how

“successful” an attack is. In a similar working paper, “Do Working Men Rebel?” Berman et al.

(2010) focus on unemployment to continue the argument regarding economic impact and suicide terrorism. The paper uses Gary Becker’s deterrence hypothesis to discuss whether the unemployed are more likely to become insurgents and actually makes the contrary assessment

(rejecting the “opportunity cost” theory), assuming that the likelihood of being an insurgent increases with employment because insurgents are more likely to be able to afford to buy information if they are employed. However, this both ignores the transnational aspect of terrorism and, as a causal analysis, ignores the underground economy’s capacity to provide means for its agents; it also focuses on violence outcomes, rather than terrorism motivations.

While these studies can motivate the inclusion of economic variables to assess terrorist trends, an economic analysis may inform whether economic trends can impact terrorism, but cannot compare the significance of those trends in relation to others, such as democratic institutions or nonviolent dispute resolution.

Exemplifying a different strain of thinking, in “Putting Women in Their Place,” Caprioli et al. (2007) make a case for focusing not on unemployment, but on gender equity, concluding that greater gender equity can lead to a reduction in violence. The analysis is focused on state security and stability, which differs from this paper’s focus on transnational terrorist groups: for example, although 15 of 19 hijackers in the 9/11 plot were Saudis, the Kingdom is relatively secure and stable, so while gender equity may also contribute to state insecurity, a more relevant measure for breeding terrorists would be to focus on the country of origin for international

terrorist perpetrators. Caprioli et al. cite an important study by M. Steven Fish, “Islam and

Authoritarianism” in which Fish asserts that gender equity may have a causal link to

democratization; however, his study aims to link mistreatment of women to autocratic Islamic

regimes, rather than terrorism in particular (Fish, 2002). This paper seeks to perform a similar

investigation to that of Fish and Caprioli et al. but with a focus on terrorist origin.

In “Sex and the Shaheed,” Bradley Thayer and Valerie Hudson (2010) discuss the life sciences reasons for a link between gender equity and suicide terrorism. While the piece references empirical studies (such as Pape’s), it does not provide its own empirical analysis;

however, it does offer a compelling argument based on analysis of empirical studies and also

provides a strong framework with which to assess the role gender equity could play in

motivating terrorist activities.

Kurrild-Klitgaard et al. examine terrorism trends and freedom, focusing on terrorist

origin similarly to this paper (Kurrild-Klitgaard et al., 2006). Kurrild-Klitgaard et al., within the

context that the Bush Administration claimed that democratizing Iraq would reduce transnational

terrorism, review much of the literature purporting that democratization is the solution to

terrorism.1 The authors also take the philosophy adopted in this paper that the country of origin

of terrorists is a commonly undervalued interest when examining the terrorist threat, “The problem with many analyses focusing solely on the amount terrorist attacks is that they do not shed any light on the characteristics of the perpetrators of terrorism, nor – in the case of transnational terrorism – on the countries from where the perpetrators originate” (Kurrild-

Klitgaard et al., 2006). The nationality focus, as well as the comparison of economic and

1 Kurrild-Klitgaard et al. direct the reader to: President Bush’s “Forward Strategy of Freedom” (announced 6 November 2003), his 2nd Inaugural Address (20 January 2005), his 2005 State of the Union speech, (5 February 2005), and, in particular, his London speech (19 November 2003). Also, see, e.g., Li, Q. (2005). “Does democracy promote or reduce transnational terrorist incidents?” Journal of Conﬂict Resolution; Rummel, R.J. (2001). “Saving lives, enriching life: Freedom as a right and a moral good.” Manuscript, http://www.mega.nu:8080/ampp/rummel/

democratic variables, make the Kurrild-Klitgaard et al. study similar to this paper. However, the

authors’ focus on “freedoms” may have created bias in their choice of variables. While, similar to this paper, they include property rights and civil liberties as compared to democracy and the economy, they stop short of examining factors that may imply that the state has policy successes, such as the benefits of law enforcement. That is, by approaching the subject from the “negative freedom” framework, the authors cannot comprehensively analyze the effectiveness of the state to prevent terrorism proclivities.2 However, Kurrild-Klitgaard et al. make a strong case that lack of democracy is a strong predictor of terrorist origin (despite statistical insignificance of the civil liberties variable and others).

If taking into account Berman et al. and Benmelech et al.’s arguments for economic indicators, Bueno de Mesquita’s argument for legal grievance, Kurrild-Klitgaard et al.’s argument for democracy, and the Caprioli et al. prioritization of women’s rights, many of the necessary indicators for assessing terrorist risk have been established individually, but not yet compared to each other. The present paper seeks to fill this empirical void. In particular, it will compare economic factors to rule of law, which could more thoroughly analyze the impact of legal grievances and women’s rights.

Rose-Ackerman suggests rule of law should receive a top priority in post-conflict reconstruction. In “Establishing the Rule of Law,” she writes that “policies that strengthen the state may simply permit those in power to act with impunity and may encourage opponents to resort to violence and threats…yet a fragile democratic state can be undermined by its own failure to limit private lawlessness” (Rose-Ackerman, 2004). Again, state failure can provide a

2 The Stanford Encyclopedia of Philosophy defines “negative liberty” as “the absence of obstacles, barriers or constraints…For example, it is sometimes said that a government should aim actively to create the conditions necessary for individuals to be self-sufficient or to achieve self-realization. The negative concept of freedom, on the other hand, is most commonly assumed in liberal defences of the constitutional liberties typical of liberal-democratic societies, such as freedom of movement, freedom of religion, and freedom of speech, and in arguments against paternalist or moralist state intervention.”

window into terrorism. Lacking rule of law leads states to fall victim to competing institutions of

organized criminals, private actors who circumvent the state for their legitimate or illegitimate

business, private economic activity without a mechanism for organization and dispute resolution,

and no means to assign responsibility and impose punishments (Rose-Ackerman, 2004). Rose-

Ackerman identifies legal institutions as fundamental to state stability in terms of criminal law

(including police and criminal process) and economic law (including property rights, contracts,

and torts) and urges that rule of law must include mechanisms for openness and accountability

(including information and auditing), as well as avenues for individual complaints (Rose-

Ackerman, 2004). Given the perception that religion or oppression are terrorists’ motivations for

their objectives, the commonality of their tactics are that they are extra-judicial, which implies

that rule of law could be a significant contributing factor to terrorist tactics and is supported by

the ideas of dispute resolution and individual rights put forth by Bueno de Mesquita and Caprioli

et al. and the state failure theory of Rose-Ackerman.

Brooks, as previously mentioned, identifies “rule of law” as a myopic new trend amongst

international development, human rights, and counterterrorism advocates alike, in her piece,

“The New Imperialism: Violence, Norms, and the ‘Rule of Law’” (Brooks, 2003). Brooks calls recent initiatives to install rule of law in at-risk states as “a string of expensive disappointments” because they promote formal legal institutions and assume, falsely, that social commitment will

follow (Brooks, 2003). This flaw, she claims, is because promotion of institutions conflates

“law” with “norms,” ignoring that law does not inherently prevent violence, but that it is

prevented when social norms lead people to buy into a legal system that punishes violence

(Brooks, 2003). Brooks identifies the central challenges to establishing meaningful rule of law

as: the priorities of choosing and justifying norms and, if necessary, changing them; and the need

to create a commitment to those norms (Brooks, 2003). Essentially, our objective in encouraging

rule of law, she claims, is only successful if our attempt at “norm entrepreneurship” supersedes that of terrorists and if the factions backing law promotion constrain those favoring extra-judicial violence (Brooks, 2003). This theory not only supports the use of outcomes as a macro-level assessment measurement, rather than institutions, but also supports examination of factors that undermine faith in a legal system, such as measuring money laundering, combat of which could reduce corruption.

In “Forming States After Failure,” Jens Meierhenrich offers some insight into how a state may convince the populace to buy into a governing system, e.g., how to engender the social norms conducive to rule of law that could supersede extra-judicial violence; “[A]gents must develop confidence in the state as a primary institutional structure…that creates opportunities for the acquisition of power, wealth and security…[by] appealing to private interests and cultivating public goods,” Meierhenrich suggests (Meierhenrich, 2004). In other words, the priority to engender buying-in to rule of law must be to create stakes for those who stand to lose from the alternative and to “facilitate the gradual construction of trust among adversaries” (Meierhenrich,

2004). Thus, it can be gleaned from Meierhenrich and Brooks that rule of law institutions cannot solve stability or terrorism problems, but that creating stakeholders for rule of law outcomes can lead to law and order, dispute resolution, and the protection of individual rights. This implies, then, that measurement of institutions cannot alone speak to trends in rule of law as a means to prevent terrorism, but only in terms of how they create stakeholders and increase rule of law outcomes.

As this review suggests, there are many competing theories on the environments conducive to terrorism. Compelling arguments have been made concerning the impact on

terrorism of economic factors, democracy, and women’s rights. However, despite attention to

governance and civil liberties, there has been little focus specifically on the outcomes of legal

regimes. Moreover, many studies test the predictability that a particular factor (such as the

economy) contributes to a terrorism-conducive environment, rather than comparing different

factors. Many studies also focus either on micro-level assessments of particular groups, or fail to test the country of origin of the terrorists, rather than their targets. And finally, others fail to integrate the factors measured with the policy-planning process. This paper seeks to fill that void by presenting a macro-level assessment of the legal, governance, and economic factors that may contribute to the risk that terrorists originate in a country.

Chapter 2. Analysis

2.1 Conceptual Framework

This paper examines the “creation” of terrorists in order to inform policy priorities in the long term to reduce the terrorist threat over time. The focus is not on demographic or psychological factors because those are difficult to measure, impossible to respond to with policy shifts, and many of them are endogenous with the structural factors herein examined. Instead, this paper focuses on many of the motivations to turn to terrorism and analyzes the social outcomes of those motivations, rather than the psychological or ethnic factors that may be associated with them. Additionally, short-term responses to terrorism require the premise that a population has already turned to terrorism, rather than preventing it, and therefore focus on effective tactics to disarm, capture or kill terrorists; this study aims to focus on preventing the social environments that breed terrorism.

Figure 1. Basic Counterterrorism Strategies Short-Term Medium-Term Long-Term

•Military: •Law Enforcement: •Diplomacy / • Capture • Arrest Development: • Kill • Deter • Influence • Prevent

Figure 2. Conceptual Framework for Terrorist Motivations

•Religious Structural Factors •Terrorism manipulation •Violence •Social angst: •Extrajudicial poverty •Weak adherence to law retribution •Social angst: lack of •Illicit activities representation •Low GDP / high unemployment •Oppression •Autocratic governance Motivations for / lack of international Potential Adverse Terrorism bargaining power Outcomes •Lack of civil liberties

In order to assess the implications of rule of law factors contributing to terrorist- producing societies, this paper examines how predictive various factors are in determining the state of origin of a terrorist perpetrator. The traditional conditions over which the U.S. may have some policy influence include economic conditions and the presence of a democratic government. Therefore, this paper measures the terrorist origin predictability of democracy, individual poverty (unemployment) and economic strain (GDP per capita), and three rule of law indicators: anti-money laundering regime strength, gender equity, and strength of lending and bankruptcy law.

Law enforcement is a necessary measure of rule of law strength, represented in this paper by the strength of a country’s anti-money laundering regime. Money laundering is a key means to organization of criminal groups. In order for illicit activity to come to fruition, and in order for those attempting to carry it out to become organized and provide themselves with expertise and resources, at least some financial transactions must take place. Additionally, money laundering investigations can act as a check on the government because those investigating the obfuscation of illicit transactions play a pivotal role in uncovering government corruption. Thus, it is

18 probable that money laundering – the attempt to obfuscate illicit transactions – is a highly significant factor in predicting criminal organization and the effectiveness of law enforcement in combating it.

Commercial due process and dispute resolution are represented in this paper by measuring the strength of lending and bankruptcy legal regimes. Strong bankruptcy laws indicate social respect for due process and also create a secure and trusted financial system that will encourage capital flows to the less prosperous (or less politically connected). Strong bankruptcy laws give confidence to financial institutions, allowing the non-wealthy to take financial risks, fostering social mobility. Aside from an indicator of a strong economy, bankruptcy laws imply order and due process within the commercial sector, which is necessary for development and a mechanism for dispute resolution. While this may overlap with some economic indicators, its implications on the rule of law are paramount due to its role in the commercial sector. Given its role in dispute resolution and commercial growth, it is also probable that states with stronger bankruptcy laws encourage commercial rule of law and therefore may mitigate terrorism.

Another central tenet of the rule of law is the rights of individuals. As discussed, Hudson and Thayer have made a strong case for the causal link between gender equity and suicide terrorism. They point out that reductions in the one-to-one ratio between males available for marriage and that of females (due to high female mortality rates or polygamy) creates a society- wide desperation for males to prove themselves worthy of a mate, causing them to act rashly to improve their reputation, which may lead to attempts at ultimate sacrifice, or suicide terrorism

(“shaheed”) to achieve ultimate social recognition. Hudson and Thayer also include anecdotal evidence that many suicide terrorists are strongly motivated by the prospect of seventy-two virgins in the afterlife if they are successful at their shaheed. Causal links aside, there is strong

evidence suggesting that those societies that have poor gender equity standards (i.e. allow for, or

even encourage, reduced rights of women compared to that of men) correspond to societies from

which terrorists originate. Protection and inclusion of women is a strong indicator of protection

of individual rights generally and therefore, given the necessity of including civil liberties in rule of law and the evidence of links to terrorism, gender equity is used in this paper to represent protection of individual rights.

By including a law enforcement indicator, a commercial due process indicator, and an indicator of individual rights, a model testing rule of law against democracy and economic indicators for predictability of terrorist origin can be comprehensive enough to draw some conclusions on the structural factors that may breed terrorism.

2.2 Data Sources

The variables of interest include the following: terrorist perpetrators by nationality, gender parity (individual rights), strength of legal rights for lending and bankruptcy (commercial law), strength of anti-money laundering regime (law enforcement), electoral process (proximity to full democracy), GDP per capita, and unemployment.

The terrorist perpetrator data used in this study is derived from the National

Counterterrorism Center’s Worldwide Incident Tracking System (WITS), limited to “perpetrator nationality” counts for incidents in the years 2007-2009 (NCTC, 2010). WITS data “consists of incidents in which subnational or clandestine groups or individuals deliberately or recklessly attacked civilians or noncombatants (including military personnel and assets outside war zones and war-like settings)” (NCTC, 2010). These are attacks initiated by terrorists that are neither foiled nor failed.

Terrorist attacks against combatants count as reckless and indiscriminate when terrorists could have reasonably foreseen that their attack would result in civilian casualties.

Therefore, combatants may be included as victims in some attacks when their presence was incidental to an attack aimed at noncombatants, and some attacks may be deemed terrorism when it recklessly affects combatants (NCTC, 2010).

NCTC codes an event as an “attack” if it results in a “premeditated kinetic event,” but near- misses are not included; for example, if a weapon is detonated but misfires, it is excluded because it did not accomplish its objective. NCTC includes such metrics as location of an attack in the theater of war or outside, as well as whether a designated foreign terrorist organization claimed responsibility for the attack. While NCTC stratifies its findings of perpetrator nationality in terms of unknown, inferred, likely, or plausible, all levels of knowledge of perpetrator nationality (greater than unknown) are included in this analysis. It should be noted that two limiting factors are particularly salient to econometric analysis with this data: 1) NCTC’s measurements include only those individuals who were caught and incident records that have

been made public, rather than conspiracies that may come to fruition at a later date or those that

remain classified; and 2) this particular analysis does not include NCTC’s listing of “unknown”

nationalities, which represents a dominant portion of the sample; but there is no reason to

conclude that the unknown nationalities, if known, would correlate differently. It is also difficult

to assess different attacks as equal, given that NCTC includes an attack resulting in zero

casualties (about one quarter of NCTC’s reported attacks) as equal to an attack resulting in

hundreds of casualties, and also includes separate attacks as part of one cohesive coordinated

attack as a single attack. However, it is extremely difficult to differentiate between attacks with

such granularity and consistency over time. Moreover, separating attacks that kill and those that

merely create terror would fail to adequately quantify the non-death costs of attacks and skew the

perception of circumstances where a certain number of people died but a very different number

could have died in similar circumstances. Another flaw is that it is difficult to separate the NCTC

data both by perpetrator nationality and by target, and by failing to separate the different types of targets, many different types of terrorists will be counted as the same, including domestic and transnational terrorists, and religiously and politically motivated.3 While it would be ideal to

stratify by victim, the limitations in the data (not least due losing many of the observations) is

significant and separation would imply that terrorist groups only attack a particular type of

victim (e.g., terrorists in Afghanistan may attack the Afghan government, the U.S. military in-

country, or migrate to another country to attack other Western targets or Americans). In addition,

as the descriptive statistics show, many countries were home to zero terrorists in a given year,

and very few were home to many terrorists; this could skew the results, but represents an

accurate view of the global terrorist threat in that a few countries are prolific producers of

terrorists while most are not at all.

3 The “sort by victim” option allows for sorting by religion, nationality, and government/police/military, but includes few options in each category and does not permit combining different categories. There is also a very limited number of observations for which these characteristics are known.

Table 1. Dependent Variable: Terrorist Perpetrators by Nationality Frequency of Terrorist Attack Perpetrators by Country of Origin Year Value 2007 2008 2009 Total 0 170 167 170 507 1-100 32 32 30 94 101-200 2 1 1 4 201-300 1 2 3 6 301-400 0 2 3 5 401-500 2 2 2 6 501-600 1 1 0 2 601-700 1 0 0 1 701-800 0 1 0 1 801-900 1 1 0 2 901-1000 0 0 1 1 Total 210 209 210 629 Source: National Counterterrorism Center

Gender equity is measured using the World Economic Forum’s (WEF) Gender Parity

Index scores for 2007-2009. The WEF measures the female-to-male ratios for economic

participation and opportunity (the participation gap, remuneration gap, and advancement gap, via labor force participation, earned income, and the number of leaders and professional

workers); educational attainment (primary, secondary, and tertiary education and literacy rates); health and survival (sex ratio at birth and life expectancy); and political empowerment (men and women in ministerial positions, parliamentary positions, and years the positions were held). A score of one indicates complete equality across all four measures, while zero represents complete

inequality. It should be noted that the diversity of factors included in this variable lends itself to a

comprehensive view of women’s rights including those unrelated to rule of law, which could

make it collinear to other variables or exaggerate its effect. However, many gender disparities

can be addressed through rule of law mechanisms, including legislation, constitutional reform, law enforcement, and dispute resolution. Moreover, a change in the gender parity variable

generally reflects a change in outcomes, rather than institution-building, and thus can show the

effectiveness of rule of law successes (e.g., greater female educational attainment can be a result

of protection of girls attending school). However, some caution is necessary because the

diversity of the gender parity variable lends itself to multicollinearity with related variables, such

as the right to vote with a ranking for democracy (see Table 8). The descriptive statistics for the

gender gap score show that no countries have a perfect score, nor a score of zero, but most

countries fall within the middle third of the rankings.

Table 2. Explanatory Variable: Individual Rights Gender Gap Index Score Year Value 2007 2008 2009 Total 0.450-0.500 1 1 1 3 0.501-0.550 1 1 2 4 0.551-0.600 14 11 10 35 0.601-0.650 29 28 24 81 0.651-0.700 53 49 49 151 0.701-0.750 22 33 39 94 0.751-0.800 5 5 5 15 0.801-0.850 3 3 4 10 Total 128 131 134 393 Source: World Economic Forum

The World Bank has assembled a score for countries’ strength of legal rights for

bankruptcy and lending, which is used in this analysis to represent commercial rule of law. The

World Bank’s index “measures the degree to which collateral and bankruptcy laws protect the

rights of borrowers and lenders and thus facilitate lending. The index ranges from 0 to 10, with

higher scores indicating that these laws are better designed to expand access to credit” (World

Bank, 2010). To compile the index, the World Bank has used data from its joint project with the

International Finance Corporation, in which the two organizations jointly keep and rank records of countries’ ease of starting a business, facilitating construction permits, registering property, getting credit, protecting investors, paying taxes, trading across borders, enforcing contracts, and closing a business. This variable’s precision is limited by the indirectness with which it is compiled, and also by the likelihood that it may overlap with the anti-money laundering regime rankings, as legal restrictions within a financial system are likely to coincide with other similar restrictions. The descriptive statistics show few countries with values of 0 or 10 and diversity of scores between the extremes.

Table 3. Explanatory Variable: Commercial Law Strength of Legal Rights for Lending and Bankruptcy Year Value 2007 2008 2009 Total 0 3 2 2 7 1 5 5 4 14 2 12 12 10 34 3 40 39 40 119 4 23 22 21 66 5 17 15 15 47 6 13 16 18 47 7 26 27 25 78 8 23 25 25 73 9 14 15 17 46 10 4 4 5 13 Total 180 182 182 544 Source: World Bank

The variable for strength of anti-money laundering regime has been compiled by the author from the State Department’s International Narcotics Control Strategy Reports (INCSR) for 2007-2009, using Volume II: Money Laundering and Financial Crimes. The State

Department has compiled a table for a large sample of countries that includes an answer of “yes”

25 or “no” regarding whether a country has criminalized money laundering, criminalized money laundering outside of narco-trafficking, records large transactions, maintains records over time, has an Egmont4-approved financial intelligence unit, has a system for identifying and enforcing the forfeiture of assets, has arrangements for asset sharing, cooperates with international law enforcement’s pursuit of money laundering investigations, controls the international transportation of currency, acquiesces to the mutual legal assistance of other states, regulates non-bank financial institutions, provides “safe harbor” for members of the financial community contributing to money laundering investigations, is party to the UN 1988 convention on narco- trafficking prevention, has criminalized the financing of terrorism, and is party to UN convention on suppressing the financing of terrorism. It also includes a “mandatory,” “possible,” or “none” option for the reporting of suspicious transactions. In order to construct the composite score, the author coded each of the options as 1 for “yes” and 0 for “no,” with the suspicious transactions reporting variable receiving a score of 0.5 for “possible” and 1 for “mandatory.” Using this score

(hereafter “INCSR Composite”) has the benefit of measuring not only laws and institutions combating financial crime, but also for the means of enforcement and the creation of stakeholders in the enforcement realm, such as reporting transactions and creating a financial intelligence unit. While it may be an oversimplification (with implications for duplicative counting) to record each variable as equal, all of the variables listed are fundamental to an anti- money laundering regime and thus any difference in how they should be weighted is debatable, if not negligible. The descriptive statistics show that many countries receive very high scores, but that there are many observations throughout all values on the scale.

4 The Egmont Group is an international forum of Financial Intelligence Units (FIU), which includes the U.S. FIU, the Financial Crimes Enforcement Network

Table 4. Explanatory Variable: Law Enforcement International Narcotics Control Strategy Report: Money Laundering Score Year Value Total 2007 2008 2009 0-1.5 4 3 3 10 2-3.5 5 6 5 16 4-5.5 6 9 5 20 6-7.5 12 10 11 33 8-9.5 21 17 14 52 10-11.5 25 17 21 63 12-13.5 24 29 29 82 14-15.5 74 74 77 225 16 21 27 28 76 Total 192 192 193 577 Source: U.S. Department of State

The democracy score used is a sub-score in Freedom House’s Freedom in the World index for 2007-2009. Freedom House’s comprehensive score was not used due to rule of law and individual rights variables that would have overlapped with other explanatory variables. The

Electoral Process score is produced by Freedom House by assessing whether the head of government or other chief national security authority is elected through free and fair elections

(established with many indicators, including politically motivated election delays, whether there is a secret ballot, equivalent weight of votes, etc.); whether national legislative representatives are elected through free and fair elections (established with many indicators, including those similar to the head-of-state evaluation); and whether electoral laws and their framework are fair

(established with many indicators, including clear and fair legislative framework, balanced election commissions, universal suffrage, and the drawing of electoral districts). One flaw in using the Freedom House variable is that it is vague and difficult to quantify. Additionally, it

may overlap with some rule of law variables, such as Freedom House’s measurement of fair

elections and the World Economic Forum’s measurement of women’s suffrage. However, the

Freedom House elections process sub-score was used because it strips out overlapping factors

that are present in Freedom House’s composite score, such as its rule of law measurement. As

such, the electoral process score represents the closest indicator of democracy with the least

overlap with rule of law. The descriptive statistics show that many countries were given perfect scores, but that there is little upward skewness, as there is still significant variance throughout the scores, and many countries were assigned zeros.

Table 5. Explanatory Variable: Similarity to Democracy Electoral Process Year Value 2007 2008 2009 Total 0 20 18 19 57 1 5 7 6 18 2 8 9 11 28 3 15 14 9 38 4 10 10 10 30 5 3 4 4 11 6 8 9 12 29 7 8 7 9 24 8 8 9 6 23 9 21 16 19 56 10 6 8 8 22 11 22 21 19 62 12 58 60 60 178 Total 192 192 192 576 Source: Freedom House

The remaining variables for this study are population and GDP per capita for 2007-2009,

compiled using World Bank data, and unemployment for 2007-2009 from the CIA World

Factbook.

Table 6. Descriptive Statistics Summary Variable Obs. Mean Std. Dev Min Max Terrorist Perpetrators 629 21.13196 97.67035 0 924 INCSR Composite 577 12.15078 3.802042 0 16 Unemployment 524 12.15078 14.11467 0 85 Electoral Process 576 7.692708 4.314306 0 12 Strength of Legal Rights 544 5.319853 2.453376 0 10 Gender Parity 393 0.672973 0.057593 0.451013 0.827603 GDP Per Capita 532 12475.99 20146.76 125.0048 183150.6

2.3 Analysis Methodology

To analyze terrorist origin countries, it is necessary to analyze the number of terrorists per country, and how intra-state conditions may correspond to shifts in the quantity of terrorists originating in a country. Using terrorist perpetrators as the dependent variable in this analysis presents a problem not easily resolved in econometric analysis: a left-censored dependent count variable that increases at an exponential rate. Econometric literature offers options for accounting for censorship as well as variables that increase exponentially, and options for analyzing count variables, but seldom are the two methods combined.

Exponential variables are typically logged to eliminate skewness and linear regression models rely on assumptions of continuity and homoskedasticity. Due to left-censorship at zero, the skewed data cannot be corrected using a logged model because a log cannot be taken for zero. Additionally, a count variable takes on non-negative integer values, is non-continuous, and is nonlinear in its parameters, therefore ruling out the appropriateness of econometric models based on assumptions of normality such as a linear regression model (Wooldridge, 2002).

Moreover, nonlinear least squares models would not account for the heteroskedasticity exhibited by count data (Wooldridge, 2002). These limitations on logged and linear models suggest it is most appropriate to use a regression model designed for count variables such as the terrorist perpetrator count examined in this paper.

Econometric literature suggests that the most appropriate model to estimate count data is the Poisson regression. Poisson relies on the conditional probabilities for values of x to obtain consistent and asymptotically normal estimators of coefficients (Wooldridge, 2002). This manner of estimating the normalized version of count data is sometimes referred to as a quasi-maximum likelihood estimation (Wooldridge, 2002).

The Poisson regression takes into account the probability that y equals the conditional value h:

P(y = h|x) = exp[-exp(xβ)][exp(xβ)]h/h!, h = 0, 1, … and constructs a log-likelihood function:

(log-likelihood)(β) = Σ li (β) = Σ {yixiβ – exp(xiβ)}, dropping the –log(yi!) term because it does not rely on β (Wooldridge, 2002).

Therefore, the basic model run in this paper is a Poisson using the following theoretical equation: 5

Terrorist Perpetrators = β0 + β1(population) + β2(GDP per capita) + β3(unemployment) + β4(electoral process score) + β5(strength score for lending and bankruptcy legal rights) + β6(gender parity index score) + β7(anti-money laundering strength score) + β8 (country dummy variables) + β9(year dummy variables)

For the purposes of this study, a Poisson regression shows the percentage effect of a change in explanatory variables on the expected changed quantity of terrorist perpetrators. One significant drawback to using the Poisson regression is that, unlike an OLS regression, it does not offer an R-squared measurement, so it is difficult to quantify the portion of the relationship between terrorist perpetrators and the explanatory variables that is explained by the model;

5 As the population across states increase, the number of perpetrators may increase simply due to the increasing size of the group, so it is necessary to control for that rise to examine the impact of structural factors. In the regression, therefore, population is included to control for its impact on other explanatory variables.

however, the Chi-squared measurement of the model denotes goodness-of-fit the way an R- squared would.

It should also be noted that unobserved effects may correlate with explanatory variables, and therefore panel data taking into account the unobserved effects was used to examine

countries over a three year period. A fixed-effects estimator is typically preferred when the

unobserved effects may correlate with the explanatory variables; however, a fixed-effects model

drops time invariant variables. Given the high number of dependent variable observations at

zero, and the consistency of some of the explanatory variables, a fixed effects model would

greatly skew this analysis, and therefore a random effects model is more appropriate.6 It should

be noted that, with terrorist perpetrator data, observations that statistically could be considered outliers must be included because those observations are the motivation for the analysis itself.7

2.4 Findings

The analysis suggests that rule of law does have an impact on terrorist perpetrator nationality.8 However, the electoral process and the strength of lending and bankruptcy legal

rights coefficients are not statistically significant. The economic conditions and other rule of law variables are, in fact, statistically significant, and therefore can be used to interpret the findings

for policy analysis.

The analysis results can be used to make interpretations for the effect of structural factors

on terrorist origin. All else equal (controlling for population), an increase in GDP per capita only

6 While a fixed-effects model may be ideal, it would drop a substantial number of observations due to time invariance. Both models were run for robustness and produced virtually identical outcomes, but the fixed-effects model dropped a large portion of the observations. 7 For the purpose of taking outliers into account, a Tobit model (latent variable model) was also run because such a model creates from data skewed due to censorship a normal, homoskedastic distribution with a linear conditional mean (Wooldridge, 2002). Thus, outliers would be included in the model without increasing the error; however, the Tobit model’s assumption that the dependent variable is continuous creates bias in the outcome of that model, so it was not used as the definitive model, but merely to explore the robustness of findings. A detailed table of the results of this model is included in the Appendix. 8 We can reject the null hypothesis that the effect of the rule of law variables is zero.

31 marginally reduces the number of expected terrorist perpetrators in a country; however, a one percent increase in the unemployment rate raises the expected number of terrorist perpetrators by

6.9 percent; a one unit increase in money laundering legal regime strength (author’s index) reduces the expected number of terrorist perpetrators by 15 percent; and a one percent increase in the gender parity index reduces the expected number of terrorist perpetrators 18.6 percent.

Table 7. Findings9 Dependent Variable: Terrorist Perpetrators by Nationality, 2007-2009 Explanatory Variables Regression Results INCSR Composite -.150***

scale: 1-16 (0.0694) Unemployment .069*** scale: 0-100 (0.0157) Electoral Process -0.006

scale: 1-12 (0.0176) Strength of Lending/ 0.023 Bankruptcy Rights scale: 1-10 (0.1478) Gender Parity -18.576***

scale: 0-1 (0.193) GDP Per Capita -0.0001*** (0.00001) # of Obs. 331 Goodness-of-fit test: Prob > Chi2 = 0.000 Standard errors are reported in parentheses *** indicates significance at the 99% level

It should also be noted that the correlation matrix implies little concern for multicollinearity, perhaps with the exception of the gender parity measurement with GDP and anti-money laundering regime strength with electoral process.10 While this certainly could skew

9 A more detailed version of this table can be found in the Appendix. As indicated in the table, the Chi-squared value indicates a strong goodness-of-fit for this model. 10 This is to be expected given the necessary overlap between women’s suffrage and economic inclusion with electoral process and economic factors, respectively.

some of the results related to those variables, it is difficult in a macro-level study to completely

separate inter-related variables such as law enforcement and protection of individual rights.

However, this could offer a partial explanation for the lack of significance of the electoral

process variable.

Table 8. Multicollinearity Test INCSR Unemplo- Electoral Strength of Gender GDP Composite yment Process Lending / Parity Per Bankruptcy Capita Correlation Matrix Legal Rights

INCSR Composite 1 Unemployment -0.2001 1 Electoral Process 0.4481 -0.0074 1 Strength of Lending / 0.3489 -0.0239 0.3201 1 Bankruptcy Legal Rights Gender Parity 0.3854 -0.2437 0.4926 0.383 1

GDP Per Capita 0.3802 -0.0556 0.3321 0.3075 0.4599 1

In order to test for robustness, a Tobit model and random effects Poisson model with terrorist perpetrators capped at 100 per country were also run to examine the impact of outliers

(see the Appendix for tables of results). However, given that the Tobit model assumes variables are continuous, and quantity of terrorist perpetrators is not, it should not be a surprise that the

Tobit model shows no significant results.11 The outlier-corrected random effects Poisson should

be questionable for external validity, given the importance of countries with high numbers of

terrorist perpetrators; however, it should be noted that the results show similar coefficients to that

of the uncorrected model, and the only variables that are statistically significant are money

laundering regime strength, strength of lending and bankruptcy rights, and unemployment. With

little statistical significance and questionable external validity, these results are discarded for the

purpose of analysis.

11 Gender parity is marginally significant; all other variables are insignificant.

2.5 Interpretation of Results

According to this analysis, economic factors appear to impact terrorism, albeit to a lesser

extent than rule of law. Interestingly, the research conflicts with findings of Berman et al. (2010), who suggest that reduced unemployment increases terrorism. The findings in this paper suggest that an increase in unemployment is expected to coincide with a marginal increase of terrorists.

However, given that the reduction in terrorist perpetrators from a reduction of unemployment is minimal, employment-driven policies may not be the most efficient in reducing terrorist perpetrators.

The rule of law variables, however, suggest that improving law enforcement and the protection of individual rights could have an impact on reducing terrorism. The strength of anti- money laundering regime result suggests that a full improvement (a country that brings a zero score to a one) in any one of the sixteen areas measured by the INCSR is expected to reduce the number of terrorist perpetrators originating from that country. For example, the creation of a financial intelligence unit (FIU) would cost a country relatively little investment, and may even generate income in the long run due to a more secure financial system.12 For a country without

an FIU to add one would imply a fifteen percent reduction in the expected quantity of terrorist

perpetrators. Additionally, many solutions to the gender parity gap can be addressed with

relatively minimal cost with a strong effect on terrorist perpetrators. For example, increasing

female enrollment in primary education could be achieved legislatively, through enforcement of

laws prohibiting preventing girls from participating in the educational process, and with public

goods such as transportation to school and the building of schools themselves. An 8.5 percent

12 “Relatively little” here is intended to act as a comparison to other effective counterterrorism efforts, such as the costs associated with military preparedness. According to the IMF, the costs of implementing an FIU vary widely, but many countries afford them with ease through the federal budget or other arrangements. For example, Belgium finances its FIU entirely through contributions from entities which benefit from it, such as stock brokers and insurance companies (IMF/WB, 2004).

increase in female primary enrollment would increase a country’s gender gap score by one

percent, implying an 18.6 percent reduction in the expected number of terrorist perpetrators to

originate from that country.13 For a government with weak gender equity, an investment in female primary enrollment can reduce the risk that terrorists develop there; and external actors can assist in the financing and technical assistance associated with new projects.

It is difficult to reach causal conclusions given these results, but the findings suggest that

there could be a strong link between rule of law and the quantity of terrorist perpetrators

originating in a country. The limitations of the Poisson model and implications of causality from

regression analysis in general imply that these results should be taken lightly, but simply that we

can reject the idea that rule of law has no impact. Moreover, the lack of reliability in the terrorist

perpetrator data as identified in the previous section and the little variance of the chosen

explanatory variables over only a three-year period call any causal implications further into question. However, given the limited availability of reliable data on terrorism, particularly before

2006, attempts to bolster internal validity (e.g., increasing the number of years of data) could reduce external validity (e.g., determining a proxy variable for comparable transnational terrorist levels in the 1980s, for example, would require a leap of logic for which there is no basis to assume such a method would provide stronger assumptions for causality).14 As such, it must be

admitted that the data used may not be able to make assumptions about a causal link between

rule of law and terrorism. However, the comparison of the rule of law variables to the economic

and other variables clearly suggests that these factors deserve the attention of policymakers in the

fields of counterterrorism and diplomacy.

13 The female net primary level enrollment ratio over the male value is weighted at 0.459 out of the educational attainment (one quarter) tranche of the total score 14 C.f., e.g., International Terrorism: Attributes of Terrorist Events, compiled by The Inter-University Consortium for Political and Social Research, offers terrorism data from 1968-1977 at no cost and 1978-2004 for a subscription; neither time-table covers information within the last several years, which brings to question its validity in a contemporary study, whereas the WITS data does not have information on incidents prior to 2004.

Chapter 3: Policy Implications and Recommendations

The findings in this paper have strong implications for counterterrorism policy priorities.

Countries that have poor rule of law in the areas examined should be considered at higher risk of being the future countries from which terrorists will commonly originate. Policies for diplomacy, development, and counterterrorism operations can use the factors identified to inform priorities, given the scarce resources they have at their disposal and difficult strategies they must engage in.

3.1 Assessing At-Risk Countries

When engaging in policy planning, there are many motivations for the U.S. government to be interested in which countries may have environments more conducive to terrorism development. The main uses of such a risk assessment can be divided into three categories: increasing situational awareness for which countries the U.S. will need to defend itself from, analyzing which countries to focus on for defensive engagement, and prioritizing where to allocate diplomatic and development resources.

In order to plan strategically to protect civilian lives and critical infrastructure, it is necessary to know the source of the threat. For the government to adequately prepare, entities such as the Department of Homeland Security must prepare risk assessments and align resources strategically. For example, if a particular country is experiencing an increase of radicalization, the immigration resources at the government’s disposal can be directed toward outreach to the immigrants from that country to reduce the probability that they will aid in the radicalization in their home country, or that they will bring it to the U.S. Another salient example was demonstrated in January 2010 when a Department of Homeland Security risk assessment was used to motivate more rigorous screening to passengers arriving from particular at-risk countries

(Trottman et al., 2010).

When a country is seen as a risk to U.S. security, the U.S. government not only prepares

for a threat against the homeland, but also seeks to mitigate the threat within the country.

Mitigating the threat can be accomplished through determining an at-risk country and providing

military aid to allies in the region or the country itself, by engaging in intelligence operations in the country, by imposing economic sanctions against the country, etc. A forward-looking risk

assessment can help prioritize which countries in which to engage and how.

Finally, when a country shows signs that, although it may or may not currently be a risk,

it may be a risk in the future, policymakers and diplomats engage the country diplomatically

through humanitarian aid, trade agreements, and regular discourse through diplomatic channels.

Given the timeline of how rule of law and other country conditions shift, it is this framework to

which a macro-level assessment such as this one would be useful.

3.2 Changing Priorities and Shifting Resources

As this paper has shown, democracy and economic development are laudable goals in

themselves and to prevent terrorism, but a driving factor in preventing terrorism in the future is

encouraging rule of law. As the literature suggests, rule of law cannot be created overnight from

scratch, but must build on existing infrastructures; simply constructing institutions will not change social norms. Instead, the literature suggests that creating stakeholders is necessary for social buy-in to rule of law. The variables used in this study were constructed with regard to stakeholders and outcomes: each of the rule of law variables not only represented criminalization and enforcement, but creation of supporting institutions and refining successful records of enforcement. Particular policies that could be derived from this paper would be: shifting funds

currently earmarked for economic development to assistance for countries with low female

educational attainment to enforce female primary enrollment, or providing technical assistance to

countries interested in establishing a financial intelligence unit. Simple reforms and resource

shifts could go a long way in preventing the next source of terrorism.

3.3 Case Study: Post-Conflict Reconstruction in Afghanistan

It is difficult to pressure a country to reform its position on rule of law. There are

stakeholders entrenched in the current system, and as such, the benefit would have to outweigh

the drawback for those in power to cede some of that power for a more dependable rule of law

system. The tactics of influencing rule of law abroad can be debated in terms of the traditional

“carrots and sticks” methods of diplomacy. However, the rebuilding process in Afghanistan

poses an example where this study could be particularly useful. The objective in Afghanistan is

to rebuild the country in such a way that terrorists will no longer originate from there, and the

infrastructure is being built from scratch. As the U.S. provides many of the resources for post-

conflict reconstruction, and plays a dominant role in technical assistance to the government, it is in a position to allocate resources to reduce the risk that Afghanistan will continue to be a breeding ground for terrorists.

3.3.1 Strategy for Rebuilding Afghanistan

As Secretary of Defense Robert Gates wrote in the introduction to the Afghanistan and

Pakistan Regional Stabilization Strategy, “the goal of the United States in Afghanistan and

Pakistan is to disrupt, dismantle, and defeat al-Qaeda and its extremist allies, and to prevent its

return to both countries” (Department of State, 2010). Secretary Gates recognized that the

objective is neither short term nor a military objective, but must be accomplished through the

long-term linking of development tactics prioritized through a lens of security concerns. He went on to write, “This goal cannot be completed through military efforts alone, but must be accompanied by the political, economic, and diplomatic efforts outlined in this plan”

(Department of State, 2010). The strategy outlines key objectives: improving governance, rule of

law, reintegration, advancing the rights of Afghan women, regional diplomacy, communications,

focused civilian assistance, and an expanded civilian presence (Department of State, 2010).

3.3.2 Safeguarding Individual Rights: Afghanistan’s Women

Soon after U.S. forces took over in Afghanistan, the U.S. highly prioritized the rights of

women:

The Americans strong-armed a handful of women into the loya jirga that then drafted the Afghan Constitution. They strong-armed a quota for women in the Afghan national legislature -- something that even American women are not treated to. U.S. troops built schools for girls and pushed for women to be included in the local shuras, and USAID ripped a page from Greg Mortenson's Three Cups of Tea and facilitated new training and educational opportunities for women. (Hudson, 2010)

Despite this initial prioritization, only twenty-seven percent of Afghan girls attend school (for fear of having their faces sprayed with acid by Talibanis) and eighty-seven percent of Afghan women report being beaten regularly (and that number is likely lower in actuality due to fear of

retribution for reporting) (Hudson, 2010). Attacks on girls’ schools continue, women continue to

be verbally or physically abused if they attempt to work, and rape is rife and scarcely enforced

(UNAMA, 2009).15 Progress in the legal regime protecting women, and enforcement of it, could

reduce these civil liberty atrocities. Given that approximately eighty percent of the population

relies on traditional dispute resolution mechanisms, which systematically fail to investigate rape,

it is unsurprising that it remains a problem.

The U.S. may sometimes shy away from “strong-arming” women’s rights because

policymakers do not wish to interfere with the social customs of the host population.

Theoretically, this is a prudent consideration, but fails to account for the rights of the women at

15 Rape itself is not a crime in Afghanistan. Chapter 8 of the Penal Code criminalizes zina, sex outside marriage, and absolves the victim of coercion; but four male witnesses must attest that a woman was coerced. The majority of women raped are between 7 and 30 years old (UNAMA, 2009).

hand and, as a policy, could be countered by taking into account the humanitarian interventions

(or failures thereof such as Rwanda in 1994) or the U.S. history of preventing tyranny of the

majority. At the least, funding for other initiatives could be withheld unless rape and attacks on

girls’ schools were criminalized with harsh penalties (and women no longer needed to provide

four witnesses to prove rape), and funding for enforcement mechanisms were increased. For

example, rape may not be reported or investigated due to corruption – and perception of corruption – throughout the law enforcement system, and instead is resolved through retributions, breeding more criminal behavior and social discontent. If the U.S. went on a communications campaign to communicate that rape and the targeting of female students were crimes, and then funded a comprehensive investigations team to focus on crimes against women, this may be a more effective means to restoring systemic order than business development.

Additionally, the U.S. could apply pressure on the Afghan government to enforce an appeals system that would allow individuals to appeal to a secular court the decisions made in informal tribunals; therefore, if a woman were convicted of extramarital intercourse when actually raped, she could appeal to a court with a more legitimate evidentiary standard.

As the literature suggests, simply creating rule of law institutions will not shift social norms, but encouraging stakeholders may aid a social shift. By creating teams of investigators, and training lawyers to prosecute crimes against women, there will be more people in

Afghanistan with a stake in reducing the oppression of women. Moreover, as women’s rights are protected, they will also become stakeholders in their own society and will no longer fear acting in their own interests. If girls are protected on their way to school, they will become educated, and the next generation of women will be prepared to take a leadership role in society. As members of the governing regime are given a stake in individual rights and women are given a

stake in their governance, Afghanistan’s social norms could shift toward those less prone to

terrorist development.

3.3.3 Enforcing the Law: Money Laundering in Afghanistan

A powerful tool in fighting terrorism is disrupting illicit means to gather and exchange resources, and the means to obfuscate those transactions. Additionally, an environment that enables financial crime is vulnerable to corruption, which can impact not only law enforcement, but governance as a whole. As this paper suggests, investment in financial crime law enforcement has a high return on investment in reducing terrorism.

Several years after the war in Afghanistan began and as commercial revitalization would be expected to take hold, “[n]ine employees of Afghanistan's Central Bank and of a troubled private bank [were] accused of stealing $1.5 million through a fake check scheme,” which led the central bank to take over Kabul Bank (Afghanistan’s largest private bank) after the removal of two top-level executives charged in the case, leading to a run on the bank (Associated Press,

2011). The resulting distrust in the banking system will undoubtedly lead to a long-term reduction in investment and donations from foreigners. The government of Afghanistan has requested that it take control of fifty percent of the foreign aid given to it, rather than being managed by foreign donors; however, it is unlikely that it can be entrusted with such responsibility, particularly in light of the International Monetary Fund (IMF) reconsidering its program in the country due to the little government reform that has taken place in the wake of the scandal (Rubin, 2011).

In 2004, the Afghan government passed anti-money laundering legislation that established the central bank’s FIU, the Financial Transactions and Reports Analysis Center of

Afghanistan (“FinTRACA”; FinTRACA, 2011). FinTRACA receives reports of suspicious

activity, analyzes those reports, and disseminates its findings to law enforcement and foreign

entitities (Sadat, 2011). FinTRACA’s investigations can not only cut down on corruption

scandals such as the one that plagued Kabul Bank, but create a check on powerful entities in

Afghanistan, engendering trust in the financial and governance systems. If there is corruption in

the government, the addition of FinTRACA diversifies stakeholders to reduce the probability

that corruption will thrive unfettered; even if FinTRACA itself is vulnerable to corruption, the

increase in corrupt entities will spread more thin the proceeds of corruption and increase the

number of individuals capable of whistle-blowing. Therefore, simply adding FinTRACA to

Afghanistan’s governance system at least creates the opportunity for corruption to be reduced.

Possibly the most significant aspect of an FIU for the purpose of this paper is its role

within the stakeholder theory presented by some of the rule of law literature. To criminalize

money laundering and establish law enforcement entities would be laudable first steps for a

rebuilding country, but an FIU plays the role of investigating those in power, both in commerce

and the government. For FinTRACA in particular, it has an independent office within the central

bank, reporting to its Executive Council (Sadat, 2011). As the country stabilizes, the commercial

sector in Afghanistan should begin to gain traction. The creation of FinTRACA creates the

possibility that, if donors become wary of corrupt entities and businesses have more of a stake in

forming licit partnerships, receiving a clean bill of health from FinTRACA will become

increasingly important for entities attempting to attract investors, which will create an incentive

for good governance.

3.3.4 Policy Lessons: Linking Short-and Long-Term Counterterrorism Objectives

What this paper suggests is that countries with better economies and stronger rule of law are less likely to produce terrorists. What this implies for policymakers is that there are policy

improvements that can impact the likelihood that terrorists will emerge in the future. An approach to foreign policy that separates development as economics and counterterrorism as a military issue must be one of the past. More focus on diplomacy and development that emphasize rule of law is paramount to counterterrorism efforts in the future.

One example of the linking of long-term counterterrorism policy with short-term diplomacy is exemplified in Section 1206 of the National Defense Authorization Act (NDAA) for Fiscal Year 2006, which provides the Secretary of Defense with the authority to “train and equip foreign military and foreign maritime security forces” only with the concurrence of the

Secretary of State (Serafino, 2011). So-called “dual-key” programs grant funding and authority to the Secretary of Defense to fund projects with security objectives, but, due to their implication on diplomacy, require the Secretary of State’s consent. A similar fund could be established for longer-term counterterrorism objectives, particularly geared toward strengthening rule of law in countries that may pose a risk of cultivating terrorists. If the Secretary of Defense was concerned that a particular country was becoming more of a risk of breeding terrorists, the State

Department, USAID, and Department of Justice could play a large role in allocating funding and implementing programs that, tactically, are development programs, but with counterterrorism

objectives.

3.4 Conclusions

As the threat of terrorism remains one of the top security concerns in the U.S., the source

of that threat is one of migrating groups, individuals, and ideas. There is no “terrorist gene,” nor

is there a single country where terrorists become radicalized or where they operate; as such, to

effectively address this concern in the long run, the U.S. must fight the conception that terrorists

are born with the objective to target civilians, or that chasing them to whichever country in

which they operate will be a sustainable strategy. In order to stop terrorists before they

radicalize, the U.S. could communicate views alternative to religious extremism, pressure for

democratic reform, or stimulate the economy of struggling countries. These may be effective

strategies, but as this paper has shown, assisting countries to strengthen adherence to the rule of

law poses a highly effective strategy to target the environments that hatch future terrorists.

As the literature suggests, simply propping up rule of law institutions cannot change

social norms, just as economic development cannot eliminate poverty, and just as democracy may elect an oppressive regime. No policies can be a panacea. However, the countries with the least likelihood of terrorists originating in them are those that have robust law enforcement regimes with many stakeholders and have high outcomes of comprehensive women’s rights. This implies that creating checks on corruption and illicit organization, such as installing anti-money laundering units, and balancing the legal system, such as allowing women to appeal a tribunal’s ruling to a secular court, could lead to long-term systemic change.

As the U.S. considers its long-term counterterrorism policy goals, preventing people from turning to terrorism begins with supporting rule of law. U.S. foreign policy must more fully integrate counterterrorism and development policy to account for the long-term.

Appendix: Regression Models

Model 1. Findings: Details Number of obs = 331 Number of groups = 120

Obs per group: min = 1 Random Effects Poisson: avg = 2.8 Terrorist Perpetrators max = 3 Wald chi2(7) = 390.21 Prob > chi2 = 0.0000 Explanatory Variable Coef. Standard z P>z 95% Conf. Interval Error INCSR Composite -.1502806 .0193077 -7.78 0.000 -.188123 -.1124381 Unemployment .0693667 .0156875 4.42 0.000 .0386197 .1001137 Electoral Process -.0059518 .0175716 -0.34 0.735 -.0403915 .028488 Strength of Lending / .0228926 .1478472 0.15 0.877 -.2668825 .3126678 Bankruptcy Legal Rights Gender Parity -18.57558 1.41259 -13.15 0.000 -21.3442 -15.80695 GDP Per Capita -.0000746 .0000134 -5.57 0.000 -.0001009 -.0000484

Model 2. Robustness Check: Tobit Number of obs = 331 2 LR chi (8) = 30.18 Tobit: Terrorist Perpetrators Prob > chi2 = .0002 Pseudo R2 = 0.023 Explanatory Variable Coef. Std. t P>|t| [95% Conf. Interval] Err. INCSR Composite -.2857463 7.171273 -0.04 0.968 -14.39388 13.82239 Unemployment .6944443 1.789626 0.39 0.698 -2.82631 4.215199 Electoral Process 2.421377 5.45977 0.44 0.658 -8.319818 13.16257 Strength of Lending / -7.968534 8.381633 -0.95 0.342 -24.45783 8.520761 Bankruptcy Legal Rights Gender Parity -704.2902 368.1579 -1.91 0.057 -1428.572 19.99156 GDP Per Capita -.0003676 .0011917 -0.31 0.758 -.0027121 .0019768 Obs. Summary: 251 left-censored (at Terrorist Perpetrators = 0) 80 uncensored 0 right-censored

Model 3. Outlier Correction Number of obs = 331

Number of groups = 120

Obs per group: min = 1 Random Effects Poisson: Terrorist Perpetrators avg = 2.8 (outlier correction) max = 3 Wald chi2(7) = 65.85 Prob > chi2 = 0 Explanatory Variable Coef. Std z P>z 95% Conf. Interval Err. INCSR Composite -.1246189 .0249859 -4.99 0.000 -.1735904 -.0756473 Unemployment -.0572768 .0171115 -3.35 0.001 -.0908148 -.0237388 Electoral Process .035557 .0272454 1.31 0.192 -.0178429 .0889569 Strength of Lending / -.293763 .1228172 -2.39 0.017 -.5344802 -.0530458 Bankruptcy Legal Rights Gender Parity -2.454253 2.123661 -1.16 0.248 -6.616553 1.708046 GDP Per Capita 0.000000842 .000021 0.04 0.968 -.0000402 .0000419

References

Associated Press. “Afghan Bankers Accused in Fake Check Scheme.” 14 February 2011.

Department of State. “Afghanistan and Pakistan Regional Stabilization Strategy.” Office of the Special Representative for Afghanistan and Pakistan. February 2010.

Benmelech, Efraim; Berrebi, Claude; Klor, Esteban F. “Economic Conditions and the Quality of Suicide Terrorism.” National Bureau of Economic Research. Working Paper 16320. August 2010.

Berman, Eli; Felter, Joseph H.; Shapiro, Jacob N.; and Callen, Michael. “Do Working Men Rebel: Insurgency and Unemployment in Afghanistan, Iraq and the Philippines.” National Bureau of Economic Research. June 2010.

Brooks, Rosa Ehrenreich. “The New Imperialism: Violence, Norms, and the ‘Rule of Law.’” Michigan Law Review. Vol. 101, No. 7. pp. 2275-2340. June 2003.

Bueno de Mesquita, Ethan. “Terrorist Factions.” Quarterly Journal of Political Science. 3: 399–418. 2008.

FinTRACA. “FinTRACA.” Financial Transactions and Reports Analysis Center of Afghanistan. Accessed 12 March 2011. http://www.fintraca.gov.af/

Fish, M. Steven. “Islam and Authoritarianism.” World Politics. Johns Hopkins University Press. 2002.

Hudson, Valerie and Leidl, Patricia. “Betrayed: They said we were there to save Afghanistan’s women. So how come we haven’t?” Foreign Policy. 10 May 2010.

Caprioli, Mary; Emmett, Chad; Hudson, Valerie M.; Ballif-Spanvill, Bonnie; McDermott, Rose. “Putting Women in Their Place.” Baker Center Journal of Applied Public Policy. Vol. 1 No. 1. Fall 2007.

Carment, David. “Preventing State Failure.” In When States Fail: Causes and Consequences. Edited by Robert I. Rotberg. Princeton University Press. 2004.

IMF / World Bank. “Financial Intelligence Units: An Overview.” International Monetary Fund, with the World Bank. 2004.

Kurrild-Klitgaard, Peter; Justesen, Mogens K.; and Klemmensen, Robert. “The Political Economy of Freedom, Democracy and Transnational Terrorism.” Public Choice. 2006.

Meierhenrich, Jens. “Forming States After Failure.” In When States Fail: Causes and Consequences. Edited by Robert I. Rotberg. Princeton University Press. 2004.

NCTC. “Methodology Utilized to Compile NCTC's Database of Terrorist Incidents.” National Counterterrorism Center. http://www.nctc.gov/witsbanner/wits_subpage_criteria.html (accessed 12 December 2010).

Pape, Robert A. “The Strategic Logic of Suicide Terrorism.” American Political Science Review. Vol. 97, No. 3 August 2003.

Rose-Ackerman, Susan. “Establishing the Rule of Law.” In When States Fail: Causes and Consequences. Edited by Robert I. Rotberg. Princeton University Press. 2004.

Rubin, Alissa. “I.M.F. Said to Have Harsh Assessment of Troubled Kabul Bank.” New York Times. 15 February 2011.

Sadat, Jafar. “Role of the Financial Intelligence Unit in Implementing an Effective Anti-Money Laundering (AML) Framework.” General Director, Financial Transactions and Reports Analysis Center of Afghanistan. Accessed 12 March 2011. http://www.google.com/url?sa=t&source=web&cd=2&ved=0CB0QFjAB&url=http%3A %2F%2Fwww.fintraca.gov.af%2Fassets%2Fppt%2FJafarSadat%2FRoleoftheFinancialIn telligenceUnitinImplementing%255B1%255D.ppt&rct=j&q=fintraca&ei=9Jp7TbC4Feux 0QGY9tDkAw&usg=AFQjCNF1ExjIeV5i6wKJU4ymr8hNuSdPwg&sig2=zXVxVuI2M noDB-WtvfuXgg&cad=rja

Serafino, Nina M. “Security Assistance Reform: ‘Section 1206’: Background and Issues for Congress.” Congressional Research Service. Report to Congress. RS22855. 11 February 2011.

Thayer, Bradley A. and Hudson, Valerie M. “Sex and the Shaheed: Insights from the Life Sciences on Islamic Suicide Terrorism.” International Security, Vol. 34, No. 4. pp. 37– 62. Spring 2010.

Trottman, Melanie and Esterl, Mike. “New Rules on Flight Security Get Airline Groups' Support.” Wall Street Journal. 5 January 2010.

UNAMA. “Silence is Violence: End the Abuse of Women in Afghanistan.” United Nations Assistance Mission in Afghanistan. 8 July 2009.

Wooldridge, Jeffrey M. “Introductory Econometrics: A Modern Approach.” South-Western College Pub. 2nd edition. pp. 540-548. 11 July 2002.

World Bank. “Strength of legal rights index.” World Bank Data. http://data.worldbank.org/indicator/IC.LGL.CRED.XQ (accessed 13 December 2010).

World Economic Forum (WEF). “The Global Gender Gap Report.” Ed. by Ricardo Hausmann, Laura D. Tyson, Saadia Zahidi. Geneva, Switzerland. 2009. https://members.weforum.org/pdf/gendergap/report2009.pdf (accessed 7 April 2011)